Energy Monitor accuracy – 20% out?

  • January 27, 2013 at 00:28 #4223
    SadGamerGeek
    Participant

    Hi,

    I’ve been using the opportunity of a newly purchased Raspberry Pi and LWRF kit to try my hand at creating some simple programs to create energy use graphs. I’d heard that some of these “clamp” type energy monitors aren’t particularly accurate so I thought I’d measure mine, and maybe build in a “calibration” value to apply to my readings to make the results more accurate.

    So, I logged my meter reading at midnight Fri night, and midnight just gone (Sat),  and compared the difference to the “Total Yesterday” logged by the LWRF Energy Monitor. My results:

    Difference in meter readings – 21.23 KWh
    Value from LWRF Monitor    – 17.15 KWh

    So, it looks like the LWRF monitor is reading around 19% too low for me. Has anyone else done this comparison? If so, is there a similar variance? Does my logic seem sound?

    I’ll try to remember to do another reading at midnight tonight (Sun) and see if the error percentage is similar. If it is, I can just incorporate a simple correction in my program and all is good. If I get a wildly different variance, not so good….

    I’m interested to hear anyone else’s thoughts on this.

    Oh, and I’ll obviously share my work with in the Hackers forum once I’m happy with it, in case it is of interest to anyone.

    Richard

    January 27, 2013 at 01:30 #4224
    nay
    Participant

    We previously had an npower supplied monitor before switching to the LWRF device and for a period dual ran and I noticed a difference which was probably at a similar level to what you have observed. Having said that I always assumed the npower device to be correct, I never verified it was though as I had no reason to doubt it.

    January 27, 2013 at 17:54 #4228
    SadGamerGeek
    Participant

    Thanks for that nay. I’m keen to hear if anyone else has done any comparisons / measurements.

    Thinking about it, when my program polls the wifi-link I notice that a reasonable number of responses come back as zero watts. My program discards those readings as they are obviously duff, but if the real readings are never being added to the total that might explain at least some of the discrepancy…..

    January 28, 2013 at 00:27 #4230
    SadGamerGeek
    Participant

    I’ve just done another comparison at midnight. 18.5% less this time, so reasonably consistent, but it does seem like quite a large margin of error.

    I’ve been wondering though, does this error apply equally to the “current usage” readings? If the “total KWh in a day” readings are out partially due to missed readings (when it returns zero), maybe the current usage ones are less affected. I need to think of a way to validate them – not so easy……

    January 28, 2013 at 09:50 #4233
    Chris
    Keymaster

    Most of the current clamp type devices are fairly inaccurate. I’m not an electrical engineer, but I understand that if the voltage isn’t exactly the same as the mathematical formula requires, the loading can’t be calculated accurately and because the clamp is battery powered and not connected to the mains, there is no way for it to measure the voltage output.

    I have dabbled in the past measuring my usage using a current cost device and bought an ‘OptiSmart’. This measures the load by reading the flashing LED on the front of my meter and therefore should be 100% accurate (or as accurate as the meter, which since my bills are based on it too, is as accurate as it’s going to get!)

    The one thing that can definitely be said is that all of these types of devices are useful for info only and shouldn’t be relied upon for accurate readings. Looking at one you can tell whether you energy usage is currently higher than your average and you can try and take steps to reduce it, but you are only ever comparing one inaccurate figure against another.

    Chris Mills
    Founder and Editor - LightwaveRF Community
    http://cpmills.com/ https://staging.lightwaverfcommunity.org.uk

    February 13, 2013 at 20:06 #4917
    SadGamerGeek
    Participant

    Just to let you know – I contacted JSJS about my findings. They said that some inaccuracy is obviously expected, but the expected variance was plus or minus 3% – drastically less than what I’m finding. They were also interested in the zero readings, so I’ve sent them a days worth of logging data (good and bad readings) and they are due to get back to me later this week.

    September 4, 2013 at 22:03 #8864
    reevemh
    Participant

    Wondered if they ever got back to you SadGamerGeek, as I’m getting some zero readings from my device as well?

    September 27, 2013 at 12:44 #9140
    Andy_EY
    Participant

    Hi there,
    I’m new here and as yet, have no LWRF products. I’m just researching the stuff and I like what I see. Really hoping the new heating controls are going to be a quality product.

    Looking around the net I found this website stating there have been some issues before with the accuracy but was corrected by an update to the Wifi link

    http://pdwhomeautomation.blogspot.co.uk/2012_03_01_archive.html

    Thought this may be of some use, of what I have read of other clamp on monitors they should be within 3%.

    Hope this is of some help to someone

    Andy

    December 12, 2013 at 18:38 #10960
    collectors
    Participant

    On a similar point, the home metering seems to be relatively ok with me, but the online chart reading is way out. At 4.30 today i was running an oven @ 2.8kw & had a 5-8 minute 9.4kw shower & none of this has showed up on the graft at all.
    Is the online part working????

    Cheers.

You must be logged in to reply to this topic.